Snap Angle Prediction for 360$^{\circ}$ Panoramas

TL;DR

This paper introduces a reinforcement learning method to predict optimal snap angles for 360-degree panoramas, improving visualization quality by reducing distortions and computational cost.

Contribution

It develops a deep recurrent neural network that learns content-aware rotations to enhance panorama projection quality.

Findings

Produces more visually pleasing panoramas

Uses 5x less computation than baseline

Effectively preserves high-level objects during rotation

Abstract

360$^{\circ}$ panoramas are a rich medium, yet notoriously difficult to visualize in the 2D image plane. We explore how intelligent rotations of a spherical image may enable content-aware projection with fewer perceptible distortions. Whereas existing approaches assume the viewpoint is fixed, intuitively some viewing angles within the sphere preserve high-level objects better than others. To discover the relationship between these optimal snap angles and the spherical panorama's content, we develop a reinforcement learning approach for the cubemap projection model. Implemented as a deep recurrent neural network, our method selects a sequence of rotation actions and receives reward for avoiding cube boundaries that overlap with important foreground objects. We show our approach creates more visually pleasing panoramas while using 5x less computation than the baseline.